Economy Intel over the next decade plans to invest 60 to 120 billion on factory space

[rob mafia]

Intel buying up TSMC's production is genius, they just cut off one of AMD's major suppliers. AMD needs fabs like TSMC and Samsung to survive, without them they have no product. And since they won't have TSMC, they have no leverage against Samsung to try to lower their price. They're now at the mercy of Samsung.

nah. i'm not even sure that's legit, but it would likely affect apple/etc more than amd. amd never uses the most cutting edge tsmc nodes (ie: apple had ~all of the 5nm), anyway. not sure why you think amd won't have tsmc - they're increasing wafers and are expected to be their 2nd biggest customer soon.

 

[jefferz]

nah. i'm not even sure that's legit, but it would likely affect apple/etc more than amd. amd never uses the most cutting edge tsmc nodes (ie: apple had ~all of the 5nm), anyway. not sure why you think amd won't have tsmc - they're increasing wafers and are expected to be their 2nd biggest customer soon.

Intel has reportedly secured the majority of TSMC's 3nm process node capacity for next year and is said to be using the technology to create three new CPUs and a GPU. If true, such a move would put pressure on rivals AMD and Apple, who rely on the semiconductor giant for their chips.
According to Chinese publication UDN, production using the 3nm node is expected to start in the second quarter of next year, with mass production set to begin in the middle of 2022. Production capacity is expected to reach 4,000 wafers by May 2022, eventually reaching 10,000 wafers per month.
The report adds that Intel will tap TSMC's 18b Fab on the 3nm process node for four of its upcoming products: three CPUs and a GPU. In the case of the former, these are non-consumer processors, likely next-gen Xeon, designed for the server and data center markets.
https://www.techspot.com/news/90780-intel-has-reportedly-secured-majority-tsmc-3nm-production.html

 

[rob mafia]

Intel has reportedly secured the majority of TSMC's 3nm process node capacity for next year and is said to be using the technology to create three new CPUs and a GPU. If true, such a move would put pressure on rivals AMD and Apple, who rely on the semiconductor giant for their chips.
According to Chinese publication UDN, production using the 3nm node is expected to start in the second quarter of next year, with mass production set to begin in the middle of 2022. Production capacity is expected to reach 4,000 wafers by May 2022, eventually reaching 10,000 wafers per month.
The report adds that Intel will tap TSMC's 18b Fab on the 3nm process node for four of its upcoming products: three CPUs and a GPU. In the case of the former, these are non-consumer processors, likely next-gen Xeon, designed for the server and data center markets.
https://www.techspot.com/news/90780-intel-has-reportedly-secured-majority-tsmc-3nm-production.html

yeah. i'm not sure that's true, but it's largely inconsequential to amd. amd wasn't planning on using 3nm in 2022. like i said, amd doesn't really use tsmc's cutting edge nodes.

zen4 is 5nm. they were never planning on having 3nm in 2022.

 

[Deorum]

@ElKarlo

“TSMC’s announced fab plan in Arizona is very vague and simply not practical,” said Mehdi Hosseini, a senior research analyst with Susquehanna International Group. “The supply chain around it has to be architected and built well in advance.”

The Fuck?

Remember when I told you this would happen years ago?

 

[ElKarlo]

@ElKarlo

“TSMC’s announced fab plan in Arizona is very vague and simply not practical,” said Mehdi Hosseini, a senior research analyst with Susquehanna International Group. “The supply chain around it has to be architected and built well in advance.”

The Fuck?

Remember when I told you this would happen years ago?

I know man. Honestly a lot of those research groups don’t know what they are talking about. They are basically just blowing anole up butts or writing fan fiction for their bosses.
Your write ups are better than the industry ones I’ve seen. TSMC is probably being vague because they want to see if they can get incentives and possibly joint ventures. JVs as in a supplier or a customer jointly builds a warehouse or a fab plant together

 

[Deorum]

I know man. Honestly a lot of those research groups don’t know what they are talking about. They are basically just blowing anole up butts or writing fan fiction for their bosses.

Your write ups are better than the industry ones I’ve seen. TSMC is probably being vague because they want to see if they can get incentives and possibly joint ventures. JVs as in a supplier or a customer jointly builds a warehouse or a fab plant together

The local industry ecosystem has been built up over the course of four decades for this moment, beginning with Intel's first flag-planting Arizona factory in 1980. There are five of them currently operational with two more under construction right now after the $20 billion in capex they just threw down in March. It's their largest manufacturing site in the world. Phoenix is as competitive with as thorough of a local supply chain as anywhere on Earth.

I also love how the TSMC saga has evolved:

"It's vague"
"It's a small offering"
"It's going to be outdated tech"

Oh, yeah?

https://arstechnica.com/gadgets/2021/05/tsmc-is-considering-a-3-nm-foundry-in-arizona/

Reuters reports that TSMC—Taiwan Semiconductor Manufacturing Company, the chip foundry making advanced processors for Apple, AMD, and Qualcomm—is beefing up its plans to build factories in Arizona while turning away from an advanced plant in Europe.

Last year, TSMC announced that it would invest $10-$12 billion to build a new 5 nm capable foundry near Phoenix, Arizona. According to Reuters' sources, TSMC officials are considering trebling the company's investment by building a $25 billion second factory capable of manufacturing 3 nm chips. More tentative plans are in the works for 2 nm foundries as the Phoenix campus grows over the next 10-15 years as well.



Like, this isn't a joke.

 

[ElKarlo]

The local industry ecosystem has been built up over the course of four decades for this moment, beginning with Intel's first flag-planting Arizona factory in 1980. There are five of them currently operational with two more under construction right now after the $20 billion in capex they just threw down in March. It's their largest manufacturing site in the world. Phoenix is as competitive with as thorough of a local supply chain as anywhere on Earth.

I also love how the TSMC saga has evolved:

"It's vague"
"It's a small offering"
"It's going to be outdated tech"

Oh, yeah?

https://arstechnica.com/gadgets/2021/05/tsmc-is-considering-a-3-nm-foundry-in-arizona/

Reuters reports that TSMC—Taiwan Semiconductor Manufacturing Company, the chip foundry making advanced processors for Apple, AMD, and Qualcomm—is beefing up its plans to build factories in Arizona while turning away from an advanced plant in Europe.

Last year, TSMC announced that it would invest $10-$12 billion to build a new 5 nm capable foundry near Phoenix, Arizona. According to Reuters' sources, TSMC officials are considering trebling the company's investment by building a $25 billion second factory capable of manufacturing 3 nm chips. More tentative plans are in the works for 2 nm foundries as the Phoenix campus grows over the next 10-15 years as well.



Like, this isn't a joke.

Dude, this is standard stuff. They pooh poohed fracking 10 years ago as fringe nonsense.
Or the opposite, like 6 or so years ago, GE under Immelt was telling us all BS, and the business media ate it all up.
It is good to be skeptical of most the news, and since you really understand the chip industry and have some inside knowledge of the Phoenix area, that makes your sources better than most of these idiots that write.
BTW putting one plant in means they will put more it. Logistics basically dictate that happens. Unless they put something somewhere nearby like New Mexico or Nevada. Which given that there is so much in Arizona, unless they get electricity discounts and other incentives, they will choose AZ.

 

[Deorum]

Dude, this is standard stuff. They pooh poohed fracking 10 years ago as fringe nonsense.

Or the opposite, like 6 or so years ago, GE under Immelt was telling us all BS, and the business media ate it all up.
It is good to be skeptical of most the news, and since you really understand the chip industry and have some inside knowledge of the Phoenix area, that makes your sources better than most of these idiots that write.

BTW putting one plant in means they will put more it. Logistics basically dictate that happens. Unless they put something somewhere nearby like New Mexico or Nevada. Which given that there is so much in Arizona, unless they get electricity discounts and other incentives, they will choose AZ.

You'd think logic would've dictated they didn't purchase 1,130 acres with the idea of throwing up a single factory and calling it a day. Intel's largest campus here is 700 ac and has the space for at least six of them. No, this is part of a massive industrial tech and advanced manufacturing park on land which was rezoned in large part for the purpose of bringing TSMC here.

 

[Deorum]

This is a really interesting topic, OP. Thank you for sharing. Not a tech person, but I appreciate the investment moves being made here. I want to learn more about the process; however, the articles I've read are like drinking from a fire hydrant to the layman.

Is there anyway to ELI5 the significance of all this for computer illiterate, dumb fucks like myself?

It's great for America's industrial high tech sector, China big joke.

 

[Clifford_Harris]

Yet another positive headline today. It's been nothing but positive headlines for several months. Things continue to look better and better for Intel.
Intel locks down all remaining TSMC 3nm production capacity, boxing out AMD and Apple

Wow, Intel is Fucked!

 

[rob mafia]

Wow, Intel is Fucked!

 

[ElKarlo]

You'd think logic would've dictated they didn't purchase 1,130 acres with the idea of throwing up a single factory and calling it a day. Intel's largest campus here is 700 ac and has the space for at least six of them. No, this is part of a massive industrial tech and advanced manufacturing park on land which was rezoned in large part for the purpose of bringing TSMC here.

That’s really dumb on their part for not even seeing that. Unless TSMC explicitly said they were getting into property development, this is the news industry being retarded and wanting to paint their own picture to suit their large cilents needs

 

[jefferz]

Here's another example of why America needs to bring manufacturing home.


ARM's "ARM China" problem has just gotten much worse
ARM China's becoming independent, without ARM's consent.

ARM's China problem has just taken a major turn for the worse, with Chinese Media reporting on Amou Technology (translation) and their plans to "promote the localization and ecological development of Arm CPU architecture" and "focus on independent research and development to create XPU products with independent architecture".
The problem with Amou Technology is that it is ARM China, ARM's rogue Chinese joint venture under Allen Wu. ARM China has rebranded itself under its rogue leader and intends to operate independently of ARM and create its own IP and products.
ARM has previously tried to oust Allen Wu from his position as CEO, but these attempts have failed so far. ARM has already cut off ARM China from its latest IP, but that's not stopping ARM China from creating its own China-focused future. ARM is currently unable to evert any control over its Chinese joint venture, and ARM employees have been forcefully blocked from entering ARM China. Now, ARM China is Amou Technologies, and this shift will re-write ARM's future.
We have previously talked about ARM's issues in China regarding their planned acquisition by Nvidia (more information here), but this move from Allen Wu/ARM China places ARM in an unprecedented situation. ARM has effectively lost control of its IP, and ARM's IP is currently set to be utilised to suit Chinese interests, not ARM's.

ARM China now serves Chinese interests
ARM China, now Amou Technologies, has started creating their own IP, independent of ARM. Amou Technologies is now using ARM's IP as a baseline to create new semiconductor IP that will be used to benefit the Chinese market. As one Chinese source puts it, "On the one hand, it acts as an agent for Arm’s CPU IP, on the other hand. Insist on local self-research and launch self-developed core power XPU architecture IP (machine translated)."
Given the nature of ARM China's, now Amou Technology's, leadership and their lack of official authorisation from ARM, it is hard to see its actions as legitimate. ARM China has gone rogue, and it now serves China's interests. How the world will react to this is anyone's guess, as ARM China has effectively stolen some of the world's most valuable semiconductor IP. What's worse is that ARM has little power to stop it...

Below is what ARM China's website currently has listed as part of its company profile (machine translated). Their plans are clear.

Relying on Arm’s world-leading ecosystem resources and technological advantages, Amou Technology (China) Co., Ltd. is based on local innovation and grows together with Chinese partners. It is committed to becoming China’s leading integrated circuit-related product core intellectual property (IP) development and The service platform supports and promotes the rapid development of China's electronic information industry.

As the only authorized operation platform of Arm's IP business in China, Amou Technology (China) Co., Ltd. will carry out integrated circuit intellectual property (IP) authorization and services to Chinese partners; and independently develop semiconductor-related IP based on the needs of the Chinese market Products, empower China's smart technology innovation.

If nothing is done about Tmou Technology, ARM will have effectively lost the rights to its IP in China. This will have ramifications on Nvidia's attempted acquisition of ARM; any IPO plans that ARM may have on the cards; and what this means for Softbank, the current owners of ARM.
The world's powers are currently fighting for control of the semiconductor market, with the EU and US seeking to manufacture more high-end semiconductor products within their territories while China attempts to grow their own semiconductor market with independent IP.
ARM China's actions will have ramifications, but we can only guess what they will be at this time.

https://www.overclock3d.net/news/cpu_mainboard/arm_s_arm_china_problem_has_just_gotten_much_worse/1

 

[Deorum]

ARM China's actions will have ramifications, but we can only guess what they will be at this time.

Is there any doubt ARM effectively transferred their IP into Chinese hands the minute they signed up for a short sighted, small profit driven bullshit joint venture? The question is what China can actually do with it. If it's anything like their x86 based processors that AMD handed the IP over for then it'll be another laughable exercise in Chinese ineptitude.









Intel is going to start wiping out all of AMD's gains over the last six quarters.

 

[rob mafia]

Is there any doubt ARM effectively transferred their IP into Chinese hands the minute they signed up for a short sighted, small profit driven bullshit joint venture?

if anything, that's still an understatement. iirc, arm/softbank sold 51% of the jv stake for a whole $775M.

the deal was monumentally stupid... which sadly, was pretty normal for softbank (ie: wework).

Intel is going to start wiping out all of AMD's gains over the last six quarters.

we'll see. i have considerable doubts still on the intel side, from validity of that to yields. and it's not like amd is standing still. seems like every time intel tries to innovate, it's an embarrassment (at worst) or negated (at best). ie: foveros and tsmc's 3d stacking/amd's lasagna.

the biggest x-factor with intel is just the giant unknown regarding management/etc. gelsinger's been trying to act like a magician, which is only adding to the intel smoke&mirrors. there's so much contradictory bullshit coming from intel that it's hard to project what their strategy even is - or if they even really have one.

ie: talks of spinning off (divesting) fabs... then a sudden 180 and idm 2.0... while then outsourcing even more to tsmc. it seems schizophrenic. it's hard to have faith in their fabs given their recent issues/bad yields (including backporting to 14nm) and it makes it even harder to trust intel to fab for you when they're outsourcing to tsmc, themselves. there's no consistency.

 

[Deorum]

if anything, that's still an understatement. iirc, arm/softbank sold 51% of the jv stake for a whole $775M.

the deal was monumentally stupid... which sadly, was pretty normal for softbank (ie: wework).



we'll see. i have considerable doubts still on the intel side, from validity of that to yields. and it's not like amd is standing still. seems like every time intel tries to innovate, it's an embarrassment (at worst) or negated (at best). ie: foveros and tsmc's 3d stacking/amd's lasagna.

the biggest x-factor with intel is just the giant unknown regarding management/etc. gelsinger's been trying to act like a magician, which is only adding to the intel smoke&mirrors. there's so much contradictory bullshit coming from intel that it's hard to project what their strategy even is - or if they even really have one.

ie: talks of spinning off (divesting) fabs... then a sudden 180 and idm 2.0... while then outsourcing even more to tsmc. it seems schizophrenic. it's hard to have faith in their fabs given their recent issues/bad yields (including backporting to 14nm) and it makes it even harder to trust intel to fab for you when they're outsourcing to tsmc, themselves. there's no consistency.

It seems they just couldn't tolerate ceding any more ground on performance while having their market share further eroded and eaten into (even if the percentages are still absurdly lopsided).

The decision to outsource to TSMC for mass production of new core products is an embarassing low point for a corp that always touted its IDM status but in exchange it allows them to compete on the strength of their microarchitecture alone, ensures no more delays on their production targets and achieves process tech parity without actually having to.

It also gives them breathing room to analyze their partially broken internal manufacturing and gear up on R&D for the next generation of EUV lithography that they - now infamously - walked away from for when their new fabs come online circa 2024-25. They won't be a player in the foundry space until at least then; it's all future bets, hopes and speculation.

I think they're going to be a massive beneficiary of geopolitics on that front with significant federal backing and they're very fortunate to be the only US based corporation the DOD could turn to in establishing a domestic foundry ecosystem. I mean, I'd love to see them get back on the horse but it matters less to me when TSMC is throwing up a half-dozen cutting edge stateside fabs.

 

[rob mafia]

It seems they just couldn't tolerate ceding any more ground on performance while having their market share further eroded and eaten into (even if the percentages are still absurdly lopsided).

The decision to outsource to TSMC for mass production of new core products is an embarassing low point for a corp that always touted its IDM status but in exchange it allows them to compete on the strength of their microarchitecture alone, ensures no more delays on their production targets and achieves process tech parity without actually having to.

It also gives them breathing room to analyze their partially broken internal manufacturing and gear up on R&D for the next generation of EUV lithography that they - now infamously - walked away from for when their new fabs come online circa 2024-25. They won't be a player in the foundry space until at least then; it's all future bets, hopes and speculation.

I think they're going to be a massive beneficiary of geopolitics on that front with significant federal backing and they're very fortunate to be the only US based corporation the DOD could turn to in establishing a domestic foundry ecosystem. I mean, I'd love to see them get back on the horse but it matters less to me when TSMC is throwing up a half-dozen cutting edge stateside fabs.

agreed, but that's kind of my point. intel doesn't seem to have a clear strategy. even if one tries to assume using tsmc is/as a stopgap, it still doesn't answer what the strategy is - if anything, it's as if they're trying to delay picking a strategy. so there's a mismash of plans... but some obvious overlaps/contradictions/conflicts of interest/etc.

and that's excluding the glofo rumours/putting zero validity to them.

 

[Arkain2K]

Samsung joining the party.

 

[ElKarlo]

Samsung joining the party.

Think Texas Instruments is building another facility in Texas as well. Good news all around